Changes in land use and land cover in

Changes in land use and land cover in Medjimurje between year 2000. and 2014. using LANDSAT satelite images SREDNJA SKOLA PRELOG Cakovecka 1 40323 Prelog Croatia Europe http: //ss-prelog. skole. hr/

Remote sensing � The proces of recording and interpreting satellite images form the air or space � LANDSAT – oldest sistem for remote sensing ( 3 active satellites )

GIS � the technology that is used for collecting and editing spatial data, databases management and for the analysis

FOSS programs �Free and open source software � QGIS i SAGA GIS �easiest data processing when creating multi- layered maps

Programs �expensive and too demanding �Arc. GIS, Global Mapper, Manifold System

Research questions 1. What are the changes in land use and land cover in Medimurje in period from year 2000. to year 2014. ? 2. What is the accuracy of classification in SAGA GIS programe? 3. Which land cover is dominating in Medimurje in year 2000 and in year 2014?

Hypotheses � only the construction class will rise over 14 years and others will remain the same or change slightly �software classification accuracy would be around 80%

Research methods � we downloaded Landsat 8 and Landsat 5 satellite images for the area of Medimurje from online site Earthexplorer § Unprocessed footage of LANDSAT 5 satellite image SAGA GIS using command Clip Grid with Polygon polygon we cut only area of Međimurje

Research methods � RGB composite is a combination of satellite images in order to make different types of images � It is possible to make a true, false and infrared combination when overlapping three satellite images

Research methods §We used infrared color combination Infrared color LANDSAT 5 satellite image as a result of command RGB Composite Infrared color LANDSAT 5 satellite image magnified on the area of water

Research methods � making shapefiles for each satellite (command Create New Shapes Layer ) � SHAPEFILE vector that appears in the form of points, lines or polygons §Infrared color LANDSAT 5 satellite image enlarged to the cell which are showing the wavelength §Making polygons at the area of the class they represent

Research methods �supervised classification has been launched and using the program we made a classification of satellite images in 5 classes that were further processed � 5 classes: water, forests, agricultural areas, construction and bare ground

Classified satellite images Classified LANDSAT 5 satellite image ( year 2000) made in SAGA GIS programe which is showing five analysed classes Classified LANDSAT 8 satellite image ( year 2014) made in SAGA GIS programe which is showing five analysed classes

Research methods �we got satellite images of LANDSAT 8 and LANDSAT 5 classified in the five classes and table for both satellites with surfaces which occupies a certain class �we have calculated the percentage of coverage for each particular class in year 2000 and 2014 �comparing percentages determining changes in the land use and land cover

Research methods �Accuracy Assesment : 1. ground truthing high costs 2. using "random points" and checks using "additional data"

Research methods �command Grid Values to Points random selection of one hundred points on the map that are later proven in QGIS programe

Research methods � in the QGIS we used Geoportal and their data as "additional data" � “additonal data" = digital orthophoto map (DOF 5) and topographic map (TK 25) � with “additional data" is used shapefile "random points" that covered the same places on map as on a classified satellite image DOF 5 map covered with shapefile "random points" TK 25 map covered with shapefile "random points"

Research methods � for each class was determined one location and the accuracy of the program classification was personally checked on the ground (Ground Truthing) The locations we visited in order to verify the accuracy of the classification

Data analysis Precentage of land cover with a specific class in Medimurje

Data analysis Classification accuracy for year 2014. §The percentage calculation in order to determine accuracy of classification in SAGA GIS programe classification accuracy 73% § Ground truthing classification accuracy 100%

Conclusion �Using FOSS programs we can make a very demanding analysis and classification of multispectral LANDSAT satellite imagery for a specific area �We assumed that the accuracy of the classification program would be about 80% but result of the study showed that the accuracy was 73%

CONCLUSION �Medjmurje is mostly agricultural area in year 2000. and in 2014. with even more than 50% agricultural coverage �from 2000. to 2014. the most changed percentage is construction with an increase of 13. 20%

Srednja skola Prelog Croatia Europe Lara, Josipa, Patricija, Valentina THANK YOU FOR ATTENTION!